<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">
<chapter>
<header>
<copyright>
<year>2004</year><year>2015</year>
<holder>Ericsson AB. All Rights Reserved.</holder>
</copyright>
<legalnotice>
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
</legalnotice>
<title>Ports and Port Drivers</title>
<prepared></prepared>
<docno></docno>
<date></date>
<rev></rev>
<file>ports.xml</file>
</header>
<p>Examples of how to use ports and port drivers are provided in
<seealso marker="doc/tutorial:introduction#interoperability tutorial">
Interoperability Tutorial</seealso>.
For information about the BIFs mentioned, see the
<seealso marker="erts:erlang">erlang(3)</seealso> manual
page in ERTS.</p>
<section>
<title>Ports</title>
<p><em>Ports</em> provide the basic mechanism for communication
with the external world, from Erlang's point of view. They
provide a byte-oriented interface to an external program. When a
port has been created, Erlang can communicate with it by sending
and receiving lists of bytes, including binaries.</p>
<p>The Erlang process creating a port is said to be
the <em>port owner</em>, or the <em>connected process</em> of
the port. All communication to and from the port must go through
the port owner. If the port owner terminates, so does the port
(and the external program, if it is written correctly).</p>
<p>The external program resides in another OS process. By default,
it reads from standard input (file descriptor 0) and writes
to standard output (file descriptor 1). The external program
is to terminate when the port is closed.</p>
</section>
<section>
<title>Port Drivers</title>
<p>It is possible to write a driver in C according to certain
principles and dynamically link it to the Erlang runtime system.
The linked-in driver looks like a port from the Erlang
programmer's point of view and is called a <em>port driver</em>.</p>
<warning>
<p>An erroneous port driver causes the entire Erlang runtime
system to leak memory, hang or crash.</p>
</warning>
<p>For information about port drivers, see the
<seealso marker="erts:erl_driver">erl_driver(4)</seealso>
manual page in ERTS,
<seealso marker="erts:driver_entry">driver_entry(1)</seealso>
manual page in ERTS, and
<seealso marker="kernel:erl_ddll">erl_ddll(3)</seealso>
manual page in Kernel.</p>
</section>
<section>
<title>Port BIFs</title>
<p>To create a port:</p>
<table>
<row>
<cell align="left" valign="middle"><c>open_port(PortName, PortSettings</c></cell>
<cell align="left" valign="middle">Returns a port identifier
<c>Port</c> as the result of opening a new Erlang port.
Messages can be sent to, and received from, a port identifier,
just like a pid. Port identifiers can also be linked to
using <c>link/1</c>, or registered under a name using
<c>register/2</c>.</cell>
</row>
<tcaption>Port Creation BIF</tcaption>
</table>
<p><c>PortName</c> is usually a tuple <c>{spawn,Command}</c>, where
the string <c>Command</c> is the name of the external program.
The external program runs outside the Erlang workspace, unless a
port driver with the name <c>Command</c> is found. If <c>Command</c>
is found, that driver is started.</p>
<p><c>PortSettings</c> is a list of settings (options) for the port.
The list typically contains at least a tuple <c>{packet,N}</c>,
which specifies that data sent between the port and the external
program are preceded by an N-byte length indicator. Valid values
for N are 1, 2, or 4. If binaries are to be used instead of lists
of bytes, the option <c>binary</c> must be included.</p>
<p>The port owner <c>Pid</c> can communicate with the port
<c>Port</c> by sending and receiving messages. (In fact, any
process can send the messages to the port, but the port owner must
be identified in the message).</p>
<p>As of Erlang/OTP R16, messages sent to ports are delivered truly
asynchronously. The underlying implementation previously
delivered messages to ports synchronously. Message passing has
however always been documented as an asynchronous operation. Hence,
this is not to be an issue for an Erlang program communicating
with ports, unless false assumptions about ports have been made.</p>
<p>In the following tables of examples, <c>Data</c> must be an I/O list.
An I/O list is a binary or a (possibly deep) list of binaries
or integers in the range 0..255:</p>
<table>
<row>
<cell align="left" valign="middle"><em>Message</em></cell>
<cell align="left" valign="middle"><em>Description</em></cell>
</row>
<row>
<cell align="left" valign="middle"><c>{Pid,{command,Data}}</c></cell>
<cell align="left" valign="middle">Sends <c>Data</c> to the port.</cell>
</row>
<row>
<cell align="left" valign="middle"><c>{Pid,close}</c></cell>
<cell align="left" valign="middle">Closes the port. Unless the
port is already closed, the port replies with
<c>{Port,closed}</c> when all buffers have been flushed
and the port really closes.</cell>
</row>
<row>
<cell align="left" valign="middle"><c>{Pid,{connect,NewPid}}</c></cell>
<cell align="left" valign="middle">Sets the port owner of
<c>Port</c>to <c>NewPid</c>. Unless the port is already closed,
the port replies with<c>{Port,connected}</c> to the old
port owner. Note that the old port owner is still linked
to the port, but the new port owner is not.</cell>
</row>
<tcaption>Messages Sent To a Port</tcaption>
</table>
<p></p>
<table>
<row>
<cell align="left" valign="middle"><em>Message</em></cell>
<cell align="left" valign="middle"><em>Description</em></cell>
</row>
<row>
<cell align="left" valign="middle"><c>{Port,{data,Data}}</c></cell>
<cell align="left" valign="middle"><c>Data</c> is received from the external program.</cell>
</row>
<row>
<cell align="left" valign="middle"><c>{Port,closed}</c></cell>
<cell align="left" valign="middle">Reply to <c>Port ! {Pid,close}</c>.</cell>
</row>
<row>
<cell align="left" valign="middle"><c>{Port,connected}</c></cell>
<cell align="left" valign="middle">Reply to <c>Port ! {Pid,{connect,NewPid}}</c>.</cell>
</row>
<row>
<cell align="left" valign="middle"><c>{'EXIT',Port,Reason}</c></cell>
<cell align="left" valign="middle">If the port has terminated for some reason.</cell>
</row>
<tcaption>Messages Received From a Port</tcaption>
</table>
<p>Instead of sending and receiving messages, there are also a
number of BIFs that can be used:</p>
<table>
<row>
<cell align="left" valign="middle"><em>Port BIF</em></cell>
<cell align="left" valign="middle"><em>Description</em></cell>
</row>
<row>
<cell align="left" valign="middle"><c>port_command(Port,Data)</c></cell>
<cell align="left" valign="middle">Sends <c>Data</c> to the port.</cell>
</row>
<row>
<cell align="left" valign="middle"><c>port_close(Port)</c></cell>
<cell align="left" valign="middle">Closes the port.</cell>
</row>
<row>
<cell align="left" valign="middle"><c>port_connect(Port,NewPid)</c></cell>
<cell align="left" valign="middle">Sets the port owner of
<c>Port</c>to <c>NewPid</c>. The old port owner <c>Pid</c>
stays linked to the port and must call <c>unlink(Port)</c>
if this is not desired.</cell>
</row>
<row>
<cell align="left" valign="middle"><c>erlang:port_info(Port,Item)</c></cell>
<cell align="left" valign="middle">Returns information as specified by <c>Item</c>.</cell>
</row>
<row>
<cell align="left" valign="middle"><c>erlang:ports()</c></cell>
<cell align="left" valign="middle">Returns a list of all ports on the current node.</cell>
</row>
<tcaption>Port BIFs</tcaption>
</table>
<p>Some additional BIFs that apply to port drivers:
<c>port_control/3</c> and <c>erlang:port_call/3</c>.</p>
</section>
</chapter>
